Multiscale Modeling of Upper Mantle Plasticity: Integrating Experimental and Theoretical data into Mean-field Schemes
Résumé
Quantifying peridotite plastic properties has been a major quest for experimental mineralogy, with direct implications for upper-mantle seismology and geodynamics. It raises, however, serious difficulties such as understanding the complex mechanisms involved within grains and at grain boundaries in multiphase aggregates deforming at high temperature (T), quantifying the effects of extreme pressures (P) on these mechanisms, and addressing stress and strain scaling issues between laboratory experiments and natural deformations. In order to address some of these issues, we developed a multiscale approach which integrates experimental deformation and diffusion data, together with first-principle calculations and theoretical considerations on mineral lattice friction (Peierls stress), within a viscoplastic self-consistent (VPSC) model for peridotite aggregates. We will present an application of a recently improved second-order (SO) VPSC scheme (e.g., Ponte Castañeda, 2002, J. Mech. Phys. Solids, 50, 737) to an olivine rich pyroxenes aggregate deformed at geological strain rate along an oceanic geotherm. Beside mineral dislocation slip systems, the SO-model extension accounts for an isotropic relaxation mechanism representing ‘diffusion-related’ creep in olivine. Slip-system critical resolved shear stresses (CRSS) are evaluated - as functions of P, T, oxygen fugacity and strain rate - from previously reported (e.g., Raterron et al., 2012, PEPI, 200-201, 105) and new experimental data (see Fraysse et al., this session), or from theoretical Peierls stress computations (e.g., Metsue et al, 2010, PCM, 37, 711). The isotropic-mechanism dependence on T and P matches that of Si self-diffusion in olivine, while its relative activity with respect to that of dislocations is constrained by reported data. The model accounts for olivine and pyroxenes known lattice preferred orientations (LPO), as well as for observed sensitivities of aggregate strength to the volume fraction of the hard phase. It shows that, to reproduce the low stresses expected in the deep upper mantle, ‘diffusion-related’ creep must contribute significantly to deformation. It also suggests that the significant weakening of olivine LPO with increasing depth results from both the P-induced [100]/[001] dislocation-slip transition and the increasing activity with T of ‘diffusion-related’ plasticity.